[Cake] Quiet

Dave Taht dave.taht at gmail.com
Wed Dec 16 11:40:09 EST 2015


I keep track of life in org-mode.

** TODO [P1] Perf
*** DONE Finish port to openwrt
**** DONE Patch 4.4

If anyone wants to play with perf for 4.4 on the linksys 1200ac, the
patches work but are sketchy.

**** NEXT Patch Kernel 3.18, 4.0, 4.1
*** Start getting valid cpu statistics back friom cake vs htb + fq_codel

* TODO Get faster delay box
Current box dies at 200mbit.
Goal: drive things past a Gbit with delay
** Faster
** Use new boxes to test with?
** Try netem with a modern kernel?
** Buy a faster box for delay
* Finish cake testing
** DONE Collect behavior of all existing models at various RTTS
*** Codel differences - nearly non-existent - and invisible with fq
http://snapon.cs.kau.se/~d/bcake_tests/

** NEXT Evaluate behavior at really low speeds
   Does the inhibit actually work?
   What does ECN do to everything else?
   capture count variable from tc_iterate
   std tests - rrul, tcp_upload, tcp_download

USE: pie, fifo, codel, cake flowblind, bcake flowblind, fq_codel, cake, bcake
VOIP test useful
*** 64k symmetric
*** at 384k/64k
*** 384k/1mbit
** Shared codel variables?
** Rip out rate estimator.
** Rip out unneeded statistics
** Substitute existing codel algorithm
** interval >> 4?
** peel at quantum
*** Other schemes
Peel at 250us backlog
Keep the device busy
Peel at some work load
Peel at quantum

** increase quantum to 2 packets at highest rates per sch_fq
** dual queue?
** Fix stats
Oct 15th notes on efficiency
> Why don't you just have a struct with all the stats, and then an array with this type of struct size TC_CAKE_MAX_TINS ?

That's how it was until a couple of days ago.  We ran out of space to
add new stats that we wanted while retaining binary backwards
compatibility with the userspace parser.  The new arrangement
specifically avoids that trap.

** Finish and Fix API
* Retest Cake
** Very low bandwidth
** Normal
** Long RTTs
** BQL
** Shaper
* Merge upstream
** Cake 2


Dave Täht
Let's go make home routers and wifi faster! With better software!
https://www.gofundme.com/savewifi


On Wed, Dec 16, 2015 at 5:35 PM, Dave Taht <dave.taht at gmail.com> wrote:
> You might be sorry you asked. ISC.org is shutting down their hosting
> so my top priority is becoming
>
> * Top 3
> ** ONGO Next year planning
> ** Unsubscribe from many mailing lists
> ** TODO [P1] bufferbloat.net move off of isc.org
> *** migrate postgres
> **** DONE Get good backup
> **** PROG Convert backup to plain text
> **** DONE DB RESTORE
> *** Migrate redmine
> *** Migrate DNS
> **** Migrate email
> **** Migrate redmine
> *** DONE Get linode site
> *** DONE Migrate site to new servers on linode
> *** Fix DNS
> **** DONE mirrors.bufferbloat.net
> *** Fix SSL
> **** Enable it properly
> *** Get new redmine conversion
> *** DONE *mostly* Merge websites together
> ** TODO [P1] Move huchra off
> *** TODO get rid of last vestige of it from DNS
> *** DONE Jenkins: discontinued
> *** DONE Remove extra files
> *** DONE Move extra files - jg + d moved
> *** NEXT Move Email
> *** Get better email stuff
> *** Move all DNS over
> **** Inspect existing tables
> *** TODO create gnugol.taht.net and cryptolisting.taht.net
> *** TODO Get certs
> *** DONE Create the-edge.taht.net on taht.net DNS
> *** DONE Move cerowrt.org over elsewhere on linode
> *** DONE Choose mailman instance 2
> *** Fix SSL
> *** Cut over
> *** Check on existing gfiber vpn clients
> *** Delete extra linodes
> *** TODO leafnode somewhere
>
> ** DONE Cerowrt.org and make-wifi-fast.org to their own servers
>
>
>
> Dave Täht
> Let's go make home routers and wifi faster! With better software!
> https://www.gofundme.com/savewifi
>
>
> On Wed, Dec 16, 2015 at 5:07 PM, Loganaden Velvindron
> <loganaden at gmail.com> wrote:
>> On Wed, Dec 16, 2015 at 4:02 PM, Kevin Darbyshire-Bryant
>> <kevin at darbyshire-bryant.me.uk> wrote:
>>> Is it quiet in here or is the cake just resting? :-)
>>>
>>
>> What's on the todo list ?
>> _______________________________________________
>> Cake mailing list
>> Cake at lists.bufferbloat.net
>> https://lists.bufferbloat.net/listinfo/cake



More information about the Cake mailing list